This project involves predicting asteroid diameters based on various orbital data and physical parameters provided by NASA JPL.
During this project, I gained extensive knowledge about analyzing high-dimensional data, fixing skewed data and removing outliers, data correlation, preprocessing pipelines, and explaining black-box models using SHAP values. Check out the report for more information.
I completed this project during my Classification and Regression (CLR204) class last year (Sep 2023).
Technologies used:
- Data analysis and engineering: NumPy and Pandas
- Data visualization and plotting: Seaborn
- Creating preprocessing pipelines and machine learning models: Scikit-Learn (might update to PyTorch later)
Dataset Link: https://github.com/blakelobato/Predicting-Asteroid-Diameter-Dash/blob/master/model/Pred_Ast_Diam_2.csv